A few thoughts about the Nikon D4
Since picking up the D4 in March it has been my go-to camera for low light and high speed photography, supplementing a D800E that's now used for just about everything else. Here are a few observations about Nikon's flagship digital SLR.
In a January interview, Nikon engineer Toshiaki Akagi described the high ISO performance of the new camera as being very similar to its predecessor, the D3S. My own subsequent comparisons of the two cameras revealed this to be true, at least when looking at NEFs that have been put through the same RAW converter on comparable conversion settings.
This means that in all but the nastiest of light, ISO 12,800 will generate acceptable image quality from the D4, just as it does with the D3S, and beyond that you get into emergency-use-only territory with both cameras. That said, you can shoot all day long at ISO 12,800 and deliver good-looking files, as long as the exposure is right and the lighting conditions aren't overwhelmingly ugly. Also, as the dance photo above illustrates, ISO 3200 is a breeze for the D4, as it is for the D3S too.
The only other RAW-related difference is an obvious one: the D4's 16.16 million pixel NEFs have slightly more real resolution than the 12.05 million pixel NEFs captured by the D3S.
In-camera JPEGs are another matter, and in this category it's a solid win for the D4 over the D3S. JPEGs that emerge directly from the D4 have a slightly crisper look as well as more natural colour in shadow areas, throughout the standard ISO range. At moderately-high ISO settings the D4's JPEG advantage is more pronounced, thanks to revised noise reduction and other processing that doesn't smudge image detail or desaturate dark colours and shadows. In fact, the D4 creates the finest in-camera JPEGs I've ever seen when set to about ISO 1600-6400 (with High ISO Noise Reduction on Normal and Sharpening, within the Neutral Picture Control, on 3). The D3S can crank out good-quality JPEGs, but the D4's are better.
The rollovers below show how D3S and D4 JPEGs differ in their handling of colour in darker areas. The ISO 12,800 comparison is a 100% magnification crop of the upper right corner of the convocation scene. Both cameras were set to Normal High ISO Noise Reduction. The D3S version is drained of colour, whereas the D4 version shows the red carpet as red.
The ISO 200 comparison shows the same thing, though the difference is more subtle at lower sensitivities. Look closely at the tree branches, the walkway and the darker portions of the building and you'll spot truer reddish colour in the D4 version.
Autofocus
The D4's AF system will be immediately familiar to anyone who has spent time with a D3S. Not just because most of the configuration options are the same, but because the autofocus feels and acts so similar too. D4 autofocus, in my experience, is very much like that of the D3S, only it's faster to acquire focus when first engaged and it's better at acquiring focus in really low light. Overall, this means the D4's AF system provides a balanced level of autofocus performance: it's good for most things, great for some things, and not horrible at anything. Some of the key traits of the D4 AF system are:
Exceptional static subject focus. The D4 consistently gets the focus distance exactly right when the subject isn't moving.
Very good, and often excellent, tracking of moving athletes in overcast conditions or under artificial light.
Fast acquisition of proper focus in sports like volleyball. The D3S feels pokey by comparison.
Good, but rarely excellent, tracking of moving athletes in full sunlight. The D4, like the D3S before it, misses a few more frames than it ought to when the light is beautiful.
The best combination of speed and tracking accuracy comes when the D4 is set to use a 9-point configuration. There are times when a single AF point makes sense, and there are situations that will benefit from all 51 AF points as well. But, like the D3S, it doesn't take much testing to figure out the D4's AF system is at its overall best when 9 AF points are active.
All of the above assumes the use of lenses like the AF-S 24-70mm f/2.8G, AF-S 70-200mm f/2.8G VR II, AF-S 400mm f/2.8G VR II and AF-S 200-400mm f/4G VR. What these all have in common is a maximum aperture of f/2.8 or slower. I've struggled to achieve the same level of autofocus performance with Nikon's f/1.4 lenses, including the AF-S 24mm f/1.4G, AF-S 50mm f/1.4G and AF-S 85mm f/1.4G. Even when the working aperture is f/2.8 and therefore shallower depth of field is not a factor.
As of this writing, I've not sorted out why f/1.4 lenses seem to bring about poorer AF system performance with the D4, or if there is a magic combination of AF settings that might alleviate the problem. With this camera, like the D3S before it, I've sidestepped the matter entirely by focusing in Live View instead. This has been a manageable workaround, at least for static subjects, but it's not much of a long term solution.
The D4's AF system can be summed up as a more responsive version of the very good AF system found in the D3S. Despite a couple of shortcomings, it's also my preferred AF system for most things. That said, I already know what I want in the D5: improved tracking in good light and more reliable autofocus with f/1.4 Nikkor glass.
Burst shooting and memory cards
Nikon has equipped the D4 with a fast CompactFlash slot, a really fast XQD slot and a large memory buffer. The data below gives an idea of how the speed of card or pair of cards you put into the camera dictates how many pictures can be shot in succession. The first number in each field represents how many frames the D4 rattled off before the buffer was full. The second number is the number of additional frames the camera was able to take after a five-second pause.
For the first table, one card was in use. For the second table, two cards were inserted and the camera was configured to either mirror all files to both cards simultaneously or write one format to the XQD card and another format to the CompactFlash card. The five cards tested represent a range of write speed potential, from really fast to moderately slow. The D4 was loaded with firmware A:1.01 B:1.01.
As you can see, the write speed capability of the card plays a big role in the number of frames that the D4 can capture in a burst. This is true, though not quite universally. With Auto Distortion Control enabled, for example, the primary factor limiting burst depth is the camera's internal processing and memory management, rather than the card. And, with both slots active, the burst depth differences between the fastest and slowest pairings become less significant, though they are still substantial.
The D4 can shoot and store a lot of pictures quickly, and a top-performing memory card can help with that, especially when only one slot is in use. If your photography requires that you blast through a lot of frames at a time, then you may want to keep Auto Distortion Control disabled (it's the only image processing-related option that so greatly impacts burst depth).
Wireless networking
I rarely head out to shoot pictures these days without an iPad, router and wireless-capable camera, so I was perhaps more excited than most when Nikon introduced Wireless Transmitter WT-5 alongside the D4 earlier this year. This dedicated Wi-Fi accessory is as small as it is expensive, but it has turned out to be money well spent. The WT-5's range and transmission speed have exceeded what I was expecting from something so tiny, its unobtrusiveness makes it practical to leave the unit attached to the camera most of the time, and it has reliably FTP'd thousands of JPEGs and NEFs to an iPad or a MacBook Air 11-inch since I got it in May.
I'd already been sending pictures wirelessly from the D4 for a couple of months before the WT-5 arrived. By connecting a Wi-Fi router to the D4's built-in Ethernet port you get all of the WT-5's key features, including the ability to file pictures to an FTP server and control the camera from a web browser. It's possible to slightly surpass the maximum wireless throughput of the WT-5 too. Plus, even a premium battery-operated router is a fraction of the cost of Nikon's wireless accessory.
The results of a round of D4 wireless speed tests are in the table below. The throughput was derived from timing how long it took to send 30 Large Fine JPEGs from the camera to the computer in FTP mode, with the wired and wireless gear described. In all tests, the separation between wireless sender and receiver for the first (or only) "wireless hop" was about 18ft/5.5m and there were no signal-blocking obstructions. For the "D4 wired" entries, an Ethernet cable linked the camera's built-in Ethernet port to the Ethernet port on the portable wired/wireless router (which sent the picture onwards, wirelessly, to the computer). Each test was performed three times and the results averaged.
The iPad throughput is an estimate of how fast ShutterSnitch's internal FTP server can take in new pictures when it doesn't also have a queue of already-received pictures to be processed for display.
For reference, the D4's built-in Ethernet can sustain a transfer rate of 5.4MB/s when sending the same batch of JPEGs to a computer across a wired Ethernet network.
The test reveals several useful things:
All three portable wired/wireless routers offer decent speed, but the d-link DAP-1350 is especially good for a unit of this type.
Given its size and low power draw, the WT-5's transmission speed is really good.
None of the options is speedy enough to handle more than moderate volumes of JPEGs, while a burst of a dozen NEFs will bog the link down for a minute or more. In other words, you have to give careful consideration to how you use the camera's networking capabilities so you can capitalize on the benefits while not overloading the connection.
As mentioned, before the WT-5 shipped I'd already converted the D4 into a wireless camera by attaching a router to it, either the Aluratek CDM530AM (because it has an internal, user-changeable battery) or d-link DAP-1350 (because it's the fastest portable router I've come across). With a short Ethernet cable running from the camera to the router, and the router in a belt pouch or backpack, you have a functional and reliable WT-5 alternative that's not overly cumbersome to use.
I covered several basketball games this way early on, where a DAP-1350 was wired to the D4 in my hand while simultaneously serving as a wireless router to a remote camera. JPEGs from both bodies streamed into ShutterSnitch on a third-generation iPad in my backpack, dependably and at a decent clip.
If the D4 Ethernet-to-router rig was working so well, you might be asking why I went ahead and shelled out for the WT-5. The answer is it provides a really enticing combination of speed and convenience. I'm a camera Wi-Fi addict, and when I discovered that the WT-5 was reliable, that its performance wasn't hobbled by its size, that I could stick it on the D4 and leave it there almost always and that I could be completely free of an Ethernet cable, I was sold. But, if the WT-5 hadn't come along, I would have been reasonably content with a portable router Ethernet-tethered to the camera.
Networking tip: If you have the WT-5 (or any other transmitter) set to ad hoc wireless networking and you're not getting the range, reliability or ease of connecting and reconnecting that you need, stop using ad hoc. Then, get yourself a router, wirelessly link the WT-5 and the destination device to it and experience a much more consistent and easy networking experience. I'd have long ago given up on sending pictures wirelessly from a camera if the only option were ad hoc.
Battery tip: Any decent battery pack with a USB output port should be able to power the d-link DAP-1350 for hours. I mainly use the Tekkeon TekCharge MP1860A, and it's good for six hours or more of semi-constant DAP-1350 activity.
Cable tip: If you'd like an Ethernet cable that exits from the camera and immediately makes a downward turn, check out networking accessories maker L-Com. Their cables feature a sturdy molded connector with a 90-degree bend and are available in black to match the D4. I have several, they're well-made and also inexpensive (though shipping outside the U.S. is not).
Either the 2ft or 3ft length is what you want if the router is to be on your belt or in your bag. The at-the-camera end of the cable should be a left-angle type (L-Com calls this "Right Angle Left Exit" in the product name) if you'd like it to bend downwards from the D4. A black 2ft cable with a standard straight connector at one end and left-angle connector at the other is here. The three foot version is here. Lengths from 1-100ft are listed here.